Photo éventuelle
Photo éventuelle
Photo éventuelle
Particulate source apportionment using two chemical transport models over French South
Eastern coastal area
Damien Piga1, Alexandre Armengaud1, Guido Pirovano2
1 AirPACA, Air Quality Observatory in Provence-Alpes-Côte-d’Azur, Marseille, France
2 RSE S.p.A, Research on Energy Systems, Milano, Italia
Introduction
The APICE project
APICE: Common Mediterranean strategy and local practical Actions for the mitigation of Port, Industries and Cities Emissions (www.apice-project.eu)
Project financed by MED 2007/2013 (from July 2010 to February 2013)
Main objective: to define local adaptation plan and common strategy to improve air quality
Introduction
The APICE project
To design efficiency actions knowledge about source contributions
Source apportionment studies
- using monitoring campaigns (PMF, CMB)
- using numerical models (CAMx, CHIMERE)
Intercomparison and evaluation
Regional monitoring network
Marseille
Introduction
Presentation of AirPACA
AirPACA: regional air quality survey in south-eastern France
• Air quality monitoring
• Air quality forecast
• Air quality information
Modeling activities:
• Emission inventory
• Daily forecast
• Scenario evaluation
Inve
nto
ry
Fore
cast
Scen
ario
Nice
Introduction
Source apportionment study
Two approaches to assess contribution of emission sources
• Receptor models → Positive Matrix Factorization (PMF), Chemical Mass Balance (CMB)
• Chemical Transport models → CHIMERE, CAMx
First step: intercomparison campaign in Marseille
with all partners participation (winter period)
Second step : long monitoring campaign in Marseille
First step: simulation of the intercomparison campaign over Marseille area using CHIMERE
Second step: set-up of CAMx model over the regional area thanks to the participation of Guido Pirovano
Third step: intercomparison of different source apportionment approaches
Data and method
The modeling system
Simulation area: 3 nested domains
• European domain (27 km) - GFR27
• Large South France domain (9 km) - FRSE9
• Regional domain (3 km) - PACA3
Anthropogenic emission data
• EMEP data (GFR27 & FRSE9)
• Local emission inventory (PACA3)
Meteorology
• WRF (GFR27 – FRSE9 – PACA3)
Natural emission data
• MEGAN (GFR27 – FRESE9 – PACA3)
Boundary and initial conditions
• Meteorological fields • GDAS - NCEP (GFR27) • WRF output (FRSE9 & PACA3)
• Chemical fields • LMDz-INCA2 (GFR27) • CHIMERE output (FRSE9 & PACA3)
Common input for CHIMERE and CAMx over PACA3 domain
Data and method
• Winter period: February 2011
• Summer period: August 2011
Name Description Color Industry – Energy Public power, heating plants, industry, waste, …
Residential – Tertiary Biomass combustion, residential plants, commercial plants, …
Natural Windblown dust, sea salts, biogenic, …
Agriculture Agriculture, forest, …
Maritime transport Shipping, loading and unloading processes, maritime activities…
Non-road transport Inland waterways, railways, air traffic, …
Road transport Cars, trucks, motorcycles, road abrasion, …
External Long-range transport from outside of the domain
The modeling system
Emission sectors involved in the source apportionment approaches
Simulation period Pollutants studied
• Focus on particles concentrations : PM10 and PM2.5
Data and method
CHIMERE model and zero-out approach
Starting equation
C°total = a.energy-industry + b.residential + c.natural + d.agriculture + e.maritime + f.non road + g.road + h.boundary conditions
Removing each emission sector, we have the following matrix system: A.X = B where X is a concentration
Contributions are given by: → contribution for industry/energy : A = a / ∑ (a, b, c, …, h)
→ contribution for residential : B = b / ∑ (a, b, c, …, h)
Reference run is used to estimate the methodology error: → Error = [ C°ref. - ∑ (X) ] / C°ref.
Data and method
CAMx model and tracer approach
9 - Titre du diaporama- 00/00/2012
Using Particulate Source Apportionment Technology (PSAT)
• Same starting equation
C°total = a.energy-industry + b.residential + c.natural + d.agriculture + e.maritime + f.non road + g.road + h.boundary conditions
• Reactive tracer methods
• Time saving (one simulation)
• Mass consistency
• Fully traceable
• Direct source apportionment PMx
• Primary particle
• Gaseous precursors
• Secondary particle
Results
CHIMERE model and zero-out approach
10 - Titre du diaporama- 00/00/2012
Daily PM10 output for the winter period at the urban background station (Marseille)
Results
CAMx model and tracer approach
11 - Titre du diaporama- 00/00/2012
Daily PM10 output for the winter period at the urban background station (Marseille)
Results
Comparison between CHIMERE and CAMx
12 - Titre du diaporama- 00/00/2012
Results at the downtown station during both winter and summer period
PM10 PM2.5
Results
Comparison between numerical models and receptor models
13 - Titre du diaporama- 00/00/2012
Results at the downtown station during the winter period
Source apportionment for primary particles only
High fraction of secondary aerosols
Good agreement between receptor models and CTMs
Main contribution from the residential sector
Results
Comparison between numerical models at the regional scale
Concentration Industry / Energy Residential Road transport
CH
IMER
E C
AM
x
Monthly PM10 output during the winter period at the regional scale
Discussion
Comparison between numerical models
15 - Titre du diaporama- 00/00/2012
Significant difference observed at the regional scale
• From CHIMERE with zero-out approach
• higher contribution from the residential sector
CHIMERE CAMx
• From CAMx with PSAT approach
• more important spatial extent for road transport and industry-energy sectors
CHIMERE CAMx
Re
side
ntial
Ro
ad In
du
stry
Due to secondary particles source apportionment
CHIMERE CAMx
Discussion
Comparison between numerical models
16 - Titre du diaporama- 00/00/2012
For the zero-out approach
• Removing a source contributing to the emissions of a gas phase precursor could have no effect on the corresponding secondary aerosol species if the removed precursor is not limiting for the conversion reaction
• Underestimation of contribution for the secondary species (as industry, road transport, …)
• Overestimation for the primary species (as biomass burning)
For the reactive tracer approach
• All sources contribute, proportionally to their weight, to secondary species, although they are in excess
• More realistic representation of contribution for the secondary species
• Higher dispersion extent for the contribution for the secondary species
Important contribution from gas phase precursor and chemistry reactions at the regional scale
Conclusion
Source apportionment over French south eastern Mediterranean coast using:
- zero-out approach with CHIMERE
- reactive tracer with CAMx
• At the large scale, significant differences between approaches due to non-linear system
- overestimation of local sources with zero-out method
- overestimation of contribution for primary emissions with zero-out method
• At the monitoring station, downtown in Marseille
- comparison between receptor models and numeric models for the primary fractions
- global good agreement during the winter period (study for the summer period in progress)
- some differences between receptor models, mainly for the industry sector
- underestimation of the natural contributions with the numerical models due to a lack for the emissions
Comparison between different approaches
Conclusion
Source apportionment study outcomes
During the winter period:
- significant contributions from industry-energy, road transport and residential sectors
During the summer period:
- significant contributions from natural emissions in a large part of the region
- inside large cities, road transport and industry-energy remain important
During the both periods:
- significant contributions of the long range transport from areas outside of the region area
Perspectives
Using these outcomes to design efficiency actions to reduce PM concentrations
Using CAMx with PSAT to apportion PM and precursors among different area
Thanks for your attention
Top Related